two ssds and 2 spining

yep... better than nothing though

I tuned my shit down quite a bit because of power use 😄

since I host it at home

@user_154092528386310154 but default timing... so 2 minute bucket refresh? and it keeps up?

how many jobs per hour?

https://discord.com/channels/821361165791133716/821361546608508938/1327194644483670027

oh right I'm blind

sorry 😄

❤️

impressive low number of failed jobs

been ok week for esi

Think I average a lot higher

Need to check

found culling unneeded keys has helped that a lot

Yeah I need to look at some of that, we generally don't delete any keys

I'm assuming mainly citadel jobs? I've been looking at them again over the holidays

That looks really nice! I'll have to have a look into this myself xD

The system pressure stat alone is really helpful for at a glance info

yeah looks that way

I really wanna make my grafana look pritty but so far I got to the “I’ve installed it” bit 😂😂

steal other peoples dashboards

and make small modifications

^

Halp 😂

https://grafana.com/grafana/dashboards/1860-node-exporter-full/

One day one day I will get a reverse proxy server running so I can run more services

thats the default linux one

Ive fought with aa for nearly 2 months and i felt exhausted after that

on the right side theres a copy id button

Ill get a look at it in a bit just finished food

import dashboard, paste id, grafana is pretty

But really liked the stats

I wanna see how my nvme raid is doing io wise on the machine

And I like nice graphs

I have dashboards.. I just don't use them, I might glance it once or twice then it becomes a "ah well"

Hehe

I have alerts and other dynamic stuff in play like discord bot editing and keeping me apprised of info etc.

Powaaahhh

which data source do i add for the linux box?

Like the color schemes!

yep

don't like the values ><

time=2025-01-10T23:41:26.278Z level=ERROR source=main.go:601 msg="Error loading config (--config.file=./prometheus.yml)" file=/home/ajdulion/prometheus-3.0.0.linux-amd64/prometheus.yml err="parsing YAML file ./prometheus.yml: yaml: line 29: did not find expected key" well rip me 😄

Is there a way to limit what some people with director/ceo roles see? I have some members in my corp, who share an alt corp, due to their alts having director role in there, they are able to see all the info of all characters between them, including those not inside the alt corp.

I realise this might be 'by design' but in our use case its not ideal

unless you changed perms

directors only get access to the corp members

not anything else

so about those citadel jobs - do you have any thoughts after looking at them again? 🙂

They have no roles at all on our SeAT instance, but they can see out of corp characters tied to the same user

Person A

• Main character - no roles in main corp

• Alt - director in alt corp

Person B

• Main Character - no roles in main corp

• Alt - Director in alt corp

both can see eachothers mains full info (and alts)

when you go to edit user on one of them

what does that screen look like

this page

Very empty lol

.docs

@user_234341298272731136, available .docs commands are: .docs.upgrade, .docs.moons, .docs.squads, .docs.donate, .docs.troubleshooting, .docs.requirements, .docs.bugs, .docs.help, .docs.auth, .docs.email, .docs.plugins, .docs.sso, .docs.install.docker, .docs.install.manual, .docs.logs

.docs.upgrade

@user_234341298272731136, Bare Metal / VM Installation Instructions: https://eveseat.github.io/docs/upgrading/from_seat_3_0/bare_metal/ Docker Installation Instructions: https://eveseat.github.io/docs/upgrading/from_seat_3_0/docker/

.docs.plugins

.docs

Hey, total rookie here at all things linux and css/html, etc. Managed to get everything working perfectly though through my own domain. How would I go about setting this to a higher setting by default?

You cant afaik

Darn, okay

w

Any idea how to manually refresh list of all corporations in eve? I cant see one corporation when trying to do squad rules.

its docker seat 5

in schedule I can't add this.

you can trigger it from the command line

If you are using the auto install script, he is already running after you enter your domain and eve's api and key.

That text is what appears after entering th eve api keys

I have a fresh install of Ubuntu server 20.04 and just added docker/docker-compose to it.

You can cd /opt/seat-docker Run docker-compose logs --tail 5 -f to see if there is any other output.

One moment, running the auto-script now to see if it gets any further...

If you're not sure, you can check to see if they're running via docker ps after you've typed in what's required.

Yeah the minute I finish entering the secret for the api key, it drops that screen.

there is a seat-docker folder in the /opt folder now.

You can start by seeing if docker is running them.

this is what's in the folder

This is what happens with docker ps

and I appreciate you taking the time to assist me.

Aside from docker and docker-compose, is there anything that needs installed? Php, etc?

I've only installed docker and docker compose. and it`s work

what server version did you install it on?

I am running Ubuntu 20.04, but I see that 24.04lts is out.

I have the latest version of docker, but I don't think veersion should affect it.

Hmm.

Can you check to see if you have .env documentation?

nano .env

Aha! That does open

(Sorry am doing this directly on the server, not in an SSH session)

Any content, or is it blank?

it's full of content

if you want to dm me, that's perfectly fine

Either your version of docker is too old or you need to install the package docker-compose-plugin

You used ``wget get.docker.com; sh index.html`` right? and when you say you installed docker-compose you mean you installed an older version of docker? I hope you didn't use snap.

no nunu was kind enough to help me get the latest version of compose

Am able to get seat running, just getting a 500 error when trying to log into it after selecting a character from eve's site.

Looks like acme isn't getting the certification.

and that's where I'm stuck

Acme shouldn't result in a 500 error, just an untrusted certificate. The 500 error I am assuming is coming from seat, so you can view it in the laravel logs

.docs.troubleshooting

https://eveseat.github.io/docs/troubleshooting/#checking-log-files

Got it to work!!!

Do you have to generate an admin login each time, or can you set a specific password?

I have a warning about adding an administrative contact, where is this done in the settings?

https://eveseat.github.io/docs/configuration/eve_administrative_contact/

You won't normally interact with it through the admin login. What is normally done is to log in with a character that is your management too (so the account is created) then logout, login as admin and modify the management user to make them an admin

Ok, so the admin is a one time thing for your main char to get set up as the "administrator"

I mean you can use it whenever, but as you've noticed it would be inconvenient to rely on using the cli to log in

yeah absolutely.

How do I fix the warning about the admnistrative contact?

Got it. Thank you all for your help on this, I really appreciate it!

Question, is the home dashboard supposed to show total mined sold, or total mined in general?

It's the value of the ore mined in total

My fork is based on the previous seat-calendar so it shouldnt rerun existing migrations... I would delete all the existing calendar tables to fix the issue

how do i fix the 500 Server error on login (after entering the website by its adress, all works fine), but still have that 500 Error

https://eveseat.github.io/docs/troubleshooting/#checking-log-files

Check to see what is causing the 500 error in the log files

[2025-01-20 03:56:32] local.ERROR: [http 404, not found] get -> https://esi.evetech.net/v1/markets/10000002/history/?datasource=tranquility&type_id=12193 [t/e: 0.08s/83] [2025-01-20 03:56:32] local.ERROR: [http 404, not found] get -> https://esi.evetech.net/v1/markets/10000002/history/?datasource=tranquility&type_id=12194 [t/e: 0.09s/82] [2025-01-20 03:56:32] local.ERROR: [http 404, not found] get -> https://esi.evetech.net/v1/markets/10000002/history/?datasource=tranquility&type_id=12195 [t/e: 0.09s/81] [2025-01-20 03:58:22] local.ERROR: [http 504, gateway timeout] get -> https://esi.evetech.net/v2/characters/2117929973/skillqueue/?datasource=tranquility [t/e: 10.05s/99] [2025-01-20 03:59:03] local.ERROR: [http 404, not found] get -> https://esi.evetech.net/v1/markets/10000002/history/?datasource=tranquility&type_id=84298 [t/e: 0.09s/99] [2025-01-20 03:59:04] local.ERROR: [http 404, not found] get -> https://esi.evetech.net/v1/markets/10000002/history/?datasource=tranquility&type_id=84300 [t/e: 0.09s/98] [2025-01-20 03:59:04] local.ERROR: [http 404, not found] get -> https://esi.evetech.net/v1/markets/10000002/history/?datasource=tranquility&type_id=84302 [t/e: 0.09s/97] [2025-01-20 03:59:05] local.ERROR: [http 404, not found] get -> https://esi.evetech.net/v1/markets/10000002/history/?datasource=tranquility&type_id=84304 [t/e: 0.06s/96] [2025-01-20 03:59:05] local.ERROR: [http 404, not found] get -> https://esi.evetech.net/v1/markets/10000002/history/?datasource=tranquility&type_id=84306 [t/e: 0.11s/95] [2025-01-20 03:59:06] local.ERROR: [http 404, not found] get -> https://esi.evetech.net/v1/markets/10000002/history/?datasource=tranquility&type_id=47115 [t/e: 0.09s/94]

but that snot the 500 error

That looks like the eseye log

you want to look in the laravel log

gimme a sec

45 /var/www/seat/vendor/laravel/framework/src/Illuminate/Console/Command.php(211): Illuminate\\Container\\Container->call()

46 /var/www/seat/vendor/symfony/console/Command/Command.php(326): Illuminate\\Console\\Command->execute()

47 /var/www/seat/vendor/laravel/framework/src/Illuminate/Console/Command.php(180): Symfony\\Component\\Console\\Command\\Command->run()

48 /var/www/seat/vendor/symfony/console/Application.php(1096): Illuminate\\Console\\Command->run()

49 /var/www/seat/vendor/symfony/console/Application.php(324): Symfony\\Component\\Console\\Application->doRunCommand()

50 /var/www/seat/vendor/symfony/console/Application.php(175): Symfony\\Component\\Console\\Application->doRun()

51 /var/www/seat/vendor/laravel/framework/src/Illuminate/Foundation/Console/Kernel.php(201): Symfony\\Component\\Console\\Application->run()

52 /var/www/seat/artisan(35): Illuminate\\Foundation\\Console\\Kernel->handle()

53 {main}

"}

The full stack trace please. If you leave the console on it then try to access the page again and it should appear there

[2025-01-20 03:17:08] local.ERROR: {"exception":"[object] (Laravel\\Socialite\\Two\\InvalidStateException(code: 0): at /var/www/seat/vendor/socialiteproviders/manager/src/OAuth2/AbstractProvider.php:50)

this makes the error

ok, mostly timeouts as i can see

so no error at all ?

The invalid state error would likely be because you used back / forward buttons while trying to login

yeah this was an old error, just going thru, and most of the errors are timeouts

can i increase timout window ?

Typically if you see a timeout in that log its not something you can or will want to increase. If you leave tail running, and try and access the page to get the 500 error, you should see that error appear in the log in real time so you can tell what the actual error is

ok, gimme asec

local.ERROR: failed to proxy request {"exception":"[object] (Seat\\Eveapi\\Exception\\TemporaryEsiOutageException(code: 500): failed to proxy request at /var/www/seat/vendor/eveseat/eveapi/src/Jobs/EsiBase.php:464)

all 500errors are proxys request errors

And it appears live when you try and access the page?

not at this moment

it happens randomly

most when the users login the first time

Ok, then it is more likely tied to jobs rather than login.. That error though sounds like something outside of seat is not happy on the TQ side

kk

how can i adjust my landing page on seat ? like other background

Do you have any jobs compoleting sucessfully?

https://discord.com/channels/821361165791133716/1247605206372323359

thx

thank you very much for your time

Anyone got any idea why someone with no roles has access to the all character info with all the information e.g. mail, assets, etc? Am I missing something?

Did you check the Access Management and make sure they are not or the permission are set correctly?

or maybe you have a open squad that has all of those permissions..

I deleted every role so no one has perimissions...

in the process of killing the docker and starting afresh to see if something went wrong somewhere

is he CEO?

nope

Just a random alt I added from a different account without linking, the character that is visible to all is director

Now I got a hostname mismatch issue... time to wait till tomorrow to try again and hope it fixes itself

So with no roles defined the following is how it should work. 1. Characters get access to all of their own data. 2. CEOs and directors get access to the data for members of their corporation

I see, in that case, if they are given a role that explicitly has no permission, would they still get access to the data for members of their corporation or would the role permissions be prioritized?

Also, can anyone point to me how to resolve a hostname conflict issue? I get the warning from browser that the connection is not private and cannot access through https. I nuked the previous vm and restarted it again from scratch, and somehow after setting it up again with the same domain, it's freaking out.

No, that permission cannot be revoked at this point in time.

I'm not sure what you mean by a hostname conflict. Can you show us what you mean?

Not sure if this is the reason, but I cannot access seat at all as I get a privacy warning and completely blocks me from accessing it

no outstanding message when I watch the docker messages except for the domain one, but that shouldn't be an issue?

uhh

Can you give me a fresh docker logs traefik? like restart fully then give the logs?

Will do after I'm back from work

.

@user_234341298272731136, available bot commands: .ping, .docs, .esi, .sde, .version

.docs

@user_234341298272731136, available .docs commands are: .docs.upgrade, .docs.moons, .docs.squads, .docs.donate, .docs.troubleshooting, .docs.requirements, .docs.bugs, .docs.help, .docs.auth, .docs.email, .docs.plugins, .docs.sso, .docs.install.docker, .docs.install.manual, .docs.logs

Unable to obtain ACME certificate for domains \"xxxxx\": unable to generate a certificate for the domains [xxxx]: acme: error: 429 :: POST :: https://acme-v02.api.letsencrypt.org/acme/new-order :: urn:ietf:params:acme:error:rateLimited :: too many certificates (5) already issued for this exact set of domains in the last 168h0m0s, retry after 2025-01-21 05:59:31 UTC: see https://letsencrypt.org/docs/rate-limits/#new-certificates-per-exact-set-of-hostnames" routerName=seat@docker rule="Host(`xxxxxx`)" providerName=primary.acme ACME CA="https://acme-v02.api.letsencrypt.org/directory"

Did you redact part of this log?

I did

removed the domain from it

Not sure why you did when it still has the domain in there

Anyways

lol, missed it

But yeah, seems like I requested too many certificates, need to wait

or is there a way to manually remove previous ones?

So on your volumes you have the letsencrypt area properly mounted and there's data in the acme.json file right?

I believe so, I used the automatic script from a new machine

You can stop traefik, remove the acme.json and do ``touch acme.json && sudo chmod 600 acme.json``

Then bring it back up but you might have to wait for the rate limits yeah..

do I have to access the traefik docker container and run those commands there?

No

Can you give your traefik volumes ?

Either way should be in one of the paths for traefik

hm... should it be listed there?

Yeah, tried deleting the acme.json and restarting it, no luck, I guess I need to wait for the rate limit...

Apparently I get one set every 36 hours, so will try again tomorrrow, thanks for the ehlp

No that's docker volumes if the acme is stored in docker volumes well that'll be a bit annoying lol.

yeah, found it in /opt/seat.../acme/

I blame you astral

Yeah just make sure to chmod 600 it or traefik will complain about it

I don't use docker volumes unless it's for data I don't care about, Acme.json is a very important file

Putting that into docker volumes isn't the best idea.

So not a me fault.

everything is yor fault 😄

Why 3 interfaces?

it also really needed it ><

You need more ram though.

(its hyper v so 2 virtual interfaces)

I wasn't counting the Vethernet

oh

Also why Hyper-V?

its a 4 port nic

But windows only seeing 3 is this windows fuckery?

(because i still had a 2022 license laying around and up untill 3 months ago i did not feel confident enough to run proxmon)

plus i used to be a windows sysop with hyper v clusters ><

dunno shows 4 in the config

if i find a cheap 10-25g built in nic i might upgrade

(not that anything needs it)

Probably would've just used a Type 1 hypervisor then ran windows in it as the Hyper-V vm's nested in windows server 2022 can be vulnerable to breakout if there's a bug with the underlying system on the Type-2 Hypervisor I assume you're doing, Vs a Hyper-V Type 1 Host or another type 1 hypervisor system as Type 2 hypervisors are generally less secure in the end regardless of what you're told.

So are you using Type 2 or Type 1 in the end?

type 1's

hyper v is still not the greates of all hypervisors

but since esxi got taken over it pritty much took over the last spot in hypervisors

Yeah not fun doing remote control on hyper-v core

oh yeah this is not a core hyper v

that shit is wack

If it's running within windows server 2022 then it's type 2 unless windows has added some fuckery?

the vm's are type 1

but the server is to old to even pass through tpm

so win 11 vm is out of the question

our storage server runs freenas for years I dread the day it needs to migrate..

very much not looking forward for that one

i'm slowly moving to docker stuff but ive been out of the game to long

So the vm's still run if the windows server os restarts?

and doign docker in windows messed me up

Docker and windows don't mix well

with this setup no there's no cluster

not to mention performance hits and you are just doing vm's in the end

if she ever gets 500+ players i might but for now I just veaam backup it all so i can migrate wherever is something crashes there's enough hardware laying around to get it back up

look in general for out storage ><

500 players?

500 eve players in seat/aa

Proxmox exists tho

and XCP-ng

yeah but i mean proxmox feels better then hyper v and esxi

im more of that esxi made itself less wanted with the new owners

omg.. dont remind me..... i have to migrate my VM to proxmox......... not looking forward of doing it...

what kind of VM?

lots of vms

also i tied XCP-ng a few years ago.. nested virtualization didnt work

a few years ago 🙂

i dont think they fixed it.. iv seen and heard proxmox works well..

hm, okay

yeah I am pretty happy with proxmox at least 🙂

which ESI is giving access to mails? I don't have the read_mail.v1 enabled on scope for the character authentication

Only have these

I have no idea how much up to date that is but might give you some answers https://skoli.ru/en/apis/eve-online/scopes

I'd assume read_mail is the one that give access to it, but on the scope of the character, it doesn't have it, but it still can see the mails...

It is esi-mail.read_mail.v1. You can take a look at which endpoint uses which scope here: https://esi.evetech.net/ui/

mail is a bit specially because there are many scopes (e.g. for sending)

So the question is, how is it reading the email of that character despite not having those scopes?

is it cached somehow or am I missing something...

it can be historical data, also every character can have different scopes depending on which login he uses

i made it so all the auth characters don't have the read_mail scope

unless it is historical, I did briefly have read_mail on that alt, then changed it so it didnt have it

any way to get rid of that cache?

relink that character and it should no longer pull new mails

what about the old mails?

they stay

do I have to delete that character on seat, or just unlink and link again? I tried that

So no way to delete the old mails from seat?

you can connect to the DB and delete them manually if you really want them gone

that's unfortunate

thank you for the info though

and by adding that plugin, this will load the calendar plugin? That's all that needs done?

basically yes.. jus remember to reload the docker container after you did that so it can apply the changes.

I shut it down, made the change, and then restarted it. Calendar showed up.

👍

Very cool, thank you. Much easier than I thought it was going to be.

Is there a calendar page that shows, like a regular calender for the month with operations?

No sure.. you could request it from the developer.

@user_301981661761896449 Apologies for the ping. Just an FYI, my Docker deployment didn't seem to like that latest eveapi update 5.0.18 (docker image v5.0.58). The new variable from your cache rework is false. It seems to lock up the character queue and spits out the attached stack trace after a while - this same job over and over, pegs out the CPU as well - did not see any other jobs throwing any errors. Did do a seat:cache:clear as well to make sure. This stack trace/error is new to this instance post-update. Originally thought it might've been due to my use of the ESEYE redis cache ESEYE_CACHE_DRIVER=redis but I have also now reproduced with the file value, so seems to be an issue in both cases.

Noted. Will have a look tonight and revert

Reverted now

However something is wrong with the arm builds which is preventing the new image coming out. Working on it.

New image ready

quick question about discord seat connector as I could not find any explanation about it on the repo, but if I install it and setup to an already existing discord server, will it reset all roles and remove it from people, or can I choose to keep everyone as they are now and choose to do a clean up at a later date after adding the bot?

Thank you! Will push it again here in a bit. It did allow me to learn how to pin the version, since with current docker configuration, composer updates all plugins on container creation via the docker entrypoint script. Originally I was just pinning the docker image version and being very confused as to why the eveapi version was always the newest - but I did manage to get stable after also pinning it in the composer.json (as a very temporary workaround) and pinning the previous docker image.

I was trying to install thought docker today since an unraid backup crashed and lost my whole install. i get stuck in a boot crash loop. If you give me a couple min i can pull the logs from the crash if i can remmeber how

I coincidentally updated my install today. Would your recommend I revert?

If on docker and you did before Crypta's revert/build being available you may want to update again - you shouldn't need to do what I did though.

How can i install a version that know will work?

Fair enough, thanks! 🙏

How can i install a specific version of seat?

if you put the "current" roles above the bot role, it wont be able to change them

nor the people that have those roles

it will how ever error every run it tries to run, and evetually hit the discord error rate limit and the job will fail

so, sort of

I have a director level Character who authed roughly 36 hours ago. But I am unable to see the Corp and/or corp information within Seat it is grey'ed out. Is there a sync that needs to run (Or I can trigger) or is there an error with that character

Actually, just general question. Which Command in Schedule syncs character data?

For most stuff, the buckets command

seat:buckets:update This guy?

Greyed out = there, but not accessible?

Correct

Yup

Are you sure you have permissions? Maybe try as admin?

Currently in as admin for Seat, let me log out and back in. Oddly enough, seat:buckets:update is set to every 2 minutes.

That is correct. The buckets job doesn’t run the updates directly, it only schedules individual character updates

So it is not permissions. In that case, how do the jobs related to his character look like?

I haven't set anything up other then Default. Where could I check jobs related to individual characters? I am mostly using Seat for structure alerts, the other Director characters (different corp)

so once I add the bot, it won't immediately just kick everyone from the current server out that is not authed to the seat server?

it never kicks anyone for any reason

it only changes roles

Yeah, only issue ever had with the bot is removing pre-set roles. But honestly I rather the bot handle it all and do it off seat. So it was not an issue. Was good

good, that's the only info I needed since it's not stated anywhere

so worst case scenario just need to give ppl roles again :p

Based on my testing, version 5.0.57 currently works fine. If you are docker to run it.

i tried bare metal and a docker install both on fresh vms that have never had any problems before. i do use cloudflare but thats easy to setup and put in the same network on docker. i have no custom dns servers that would interfere with downloading anything. it just wont work.

The latest docker image should be fine again, It was only 5.0.58- 5.0.60 that had issues. The change was removed from 5.0.61

Can confirm 5.0.61 is working fine.

I had this problem when installing 5.0.61

scheduler-2-1 | Script @php artisan package:discover --ansi handling the post-autoload-dump event returned with error code 1 front-2-1 | front-2-1 | In Connection.php line 829: front-2-1 | front-2-1 | SQLSTATE[42S02]: Base table or view not found: 1146 Table 'seat2.globalset front-2-1 | tings' doesn't exist (Connection: mysql, SQL: select value from `globals front-2-1 | ettings where name` = admincontact limit 1) front-2-1 | front-2-1 | front-2-1 | In Connection.php line 423: front-2-1 | front-2-1 | SQLSTATE[42S02]: Base table or view not found: 1146 Table 'seat2.globalset front-2-1 | tings' doesn't exist

Agreed on the installation method, which was not available before 5.0.57. It's all docker script installation.

At the top, there should be a truck item. Click it. Then, go to "Failed Jobs". If a job for a character fails, it should appear there. You can search for the character id of the character in question so see if his jobs fail. If there is too much noise, you can go to the "Monitor" and use the character id as monitoring tag. Let it run for an hour or two and you'll see all jobs, failed or not, related to that character.

Did you change the database name to seat2?

I'm running two seat containers on docker, so changed some of the container and data names.

🫠 I think I know what the problem is, I'll retry it.

Still the same problem, the indicator name changed from seat2 to seat🫠 。 scheduler-1 | In Connection.php line 829: scheduler-1 | scheduler-1 | SQLSTATE[42S02]: Base table or view not found: 1146 Table 'seat.globalsett scheduler-1 | ings' doesn't exist (Connection: mysql, SQL: select value from `globalse scheduler-1 | ttings where name` = admincontact limit 1) scheduler-1 | scheduler-1 | scheduler-1 | In Connection.php line 423: scheduler-1 | scheduler-1 | SQLSTATE[42S02]: Base table or view not found: 1146 Table 'seat.globalsett scheduler-1 | ings' doesn't exist

This looks oddly familiar to the error I had yesterday

Let me check my log files

Interesting, I would have expected it briefly in containers other than front on first boot. But not in front... Well that makes a design decision for me. I was always tossing up setting the user agent for eseye on boot, or on each request. I went with boot due to the ever so slight potential for performance impact by doing it every time. I will switch it to instead set it on request

This was the fix suggested to me yesterday

That works to get you going for now, but means you will no longer receive updates to core packages because you have pinned to that version

yea my hope is what ever borked things will be resolved soon so i can go back to getting updates

i had done a fresh install on the 5th of jan and had zero issues on the default install through docker so the issues had to have happened in between now and then

I will have a patch up shortly to address the issue :). I will also run the install script to make sure it works once pushed

It is 100% a first world problem, but I am -1 monitor and its really putting me off with my normal workflow and slowing me down

@user_301981661761896449 we believe in you!

oof

this is no joke

On the plus side, by changing to setting UA each time, I can also revert the warning about restarting workers when changing admin contact

after update my server is not working now with this error

seat-docker-front-1 | > Illuminate\Foundation\ComposerScripts::postAutoloadDump
seat-docker-front-1 | Script Illuminate\Foundation\ComposerScripts::postAutoloadDump handling the post-autoload-dump event terminated with an exception
seat-docker-front-1 |
seat-docker-front-1 | In platform_check.php line 26:
seat-docker-front-1 |
seat-docker-front-1 |   Composer detected issues in your platform: Your Composer dependencies requi
seat-docker-front-1 |   re a 64-bit build of PHP.
seat-docker-front-1 |
seat-docker-front-1 |
seat-docker-front-1 | dump-autoload [-o|--optimize] [-a|--classmap-authoritative] [--apcu] [--apcu-prefix APCU-PREFIX] [--dry-run] [--dev] [--no-dev] [--ignore-platform-req IGNORE-PLATFORM-REQ] [--ignore-platform-reqs] [--strict-psr] [--strict-ambiguous]
seat-docker-front-1 |

any solution?

are you on a 32 bit system?

ooo@instance-20240517-2309:/opt/seat-docker$ uname -na
Linux instance-20240517-2309 6.8.0-1018-oracle #19-Ubuntu SMP Mon Dec  9 21:00:17 UTC 2024 aarch64 aarch64 aarch64 GNU/Linux

That looks like an oracle instance maybe or Arm64

yeah it is

aarch64 is 64 bit. hmm

I am not sure why it shows this message after update

before upgrade it was working well

We did have some issues with arm docker builds recently and had to change some stuff in the build process. What is your current seat-docker version? If it is up-to-date, could you try downgrading to seat-docker 5.0.57 to see if that fixes the issue?

should i fix with it?

yes, I'd like to see if that fixes the issue

same problem

seat-docker-worker-1 | starting workers via horizon
seat-docker-worker-1 | Processing plugins from SEAT_PLUGINS
seat-docker-worker-1 | Completed plugins processing
seat-docker-worker-1 | Dumping the autoloader
seat-docker-worker-1 | Generating optimized autoload files
seat-docker-worker-1 | > Illuminate\Foundation\ComposerScripts::postAutoloadDump
seat-docker-worker-1 | Script Illuminate\Foundation\ComposerScripts::postAutoloadDump handling the post-autoload-dump event terminated with an exception
seat-docker-worker-1 |
seat-docker-worker-1 | In platform_check.php line 26:
seat-docker-worker-1 |
seat-docker-worker-1 |   Composer detected issues in your platform: Your Composer dependencies requi
seat-docker-worker-1 |   re a 64-bit build of PHP.
seat-docker-worker-1 |
seat-docker-worker-1 |
seat-docker-worker-1 | dump-autoload [-o|--optimize] [-a|--classmap-authoritative] [--apcu] [--apcu-prefix APCU-PREFIX] [--dry-run] [--dev] [--no-dev] [--ignore-platform-req IGNORE-PLATFORM-REQ] [--ignore-platform-reqs] [--strict-psr] [--strict-ambiguous]
seat-docker-worker-1 |
seat-docker-worker-1 exited with code 1

did you change it for all containers or just front?

just front

is front fine?

oh yeah

I will fix othere things also

let me check

It works well now. Thanks!

Great. Just remember to remove the hardcoded version again once the issue is fixed, as otherwise you won't get new updates

Got problems with no ESI loading and no chardata updating. But diagnose says all green. What can i check next?

I added a secondary account with Director and it loaded up within 10 minutes. However, I also checked and saw another account was generating errors for corp sync in the spot you listed above. I removed that account in that same 10 minutes. One of the two fixed it.

Hey all, ran into this problem after a docker install, went and ran it manually and same thing.... Anyone seen this before? ´´scheduler-1 exited with code 1 front-1 | front-1 | In Connection.php line 829: front-1 | front-1 | SQLSTATE[HY000] [1045] Access denied for user 'seat'@'172.23.0.3' (using pa front-1 | ssword: YES) (Connection: mysql, SQL: select value from global_settings front-1 | where name = admin_contact limit 1) front-1 | front-1 | front-1 | In Connector.php line 65: front-1 | front-1 | SQLSTATE[HY000] [1045] Access denied for user 'seat'@'172.23.0.3' (using pa front-1 | ssword: YES) front-1 | front-1 | front-1 | Script @php artisan package:discover --ansi handling the post-autoload-dump event returned with error code 1´´

Update again

It was fixed in the last docker builds.

Unless you are on an arm based machine, in which case please pin to docker version 5.0.57

@user_301981661761896449 Im on an 86x64, and have tried both ideas for good measure

Use the docker compose pull command to make sure you have the latest images then up -d to apply them. If you are still seeing the access denied error then docker compose down -v and then up -d this will wipe the db and recreate it. Noting to use the full docker compose command as per docs

Ya I eneded up removing the stack and restarted and it worked so pretty sure it was the db needing recreated. But now I've ran into a bit of another problem: After signing into seat as a normal user it returns error 500 saying the token has expired

https is also not working when it had before 🤷‍♂️ Wanna say thank you for any help too 🙂

If you wiped the db you wiped all of the seat in the end.

ya just not sure what the problem is now

Is there a guide or Checklis, what i can check if Charcters and all Items arent Updating ? all green on my side with the diagnose command, but still no Updates. Please.....

Found an not working supervisor. he is missing the the Supervisor Version 3.1.1 but dpkg says 3.1.1-1.1 is installed. Should be fine but keep crashing. Maybe thats an Error and how can i fixed that ?

Found it. Had to chnage the Phyton Version in the supervisord file to version 2 in the Shebang first line. Strange

So any ideas?

two things. Did the person sit on the CCP SSO login page for a long time before clicking continue? Also is your server time set correctly?

It was the server time, thank you very much!

first time doin this on a friends server in Korea XD

Sometimes it helps to clear unused volumes but in my case that didn’t do much .. now I have enough storage so I didn’t really bother to look any further

Not sure why reinstalling would be an approach to take

There are a couple of things you can do to try and reclaim some space 1. to cleanup unused images and volume you can use the command docker system prune NOTE

WARNING! This will remove:
  - all stopped containers
  - all networks not used by at least one container
  - all dangling images
  - unused build cache

2. Make sure the maintenance job is running

You can enable cleanup stale data if you are happy to lose old data for toons you no longer have tokens for

But that said, 54GB for 1500ish toons is not overly unreasonable either. Particularly with an accumulated history. It can easily be larger if you had a few prolific market traders for example

As to your questions 1. Yes so long as you backup everything and do not change the eve developer application 2. Maybe, maybe not 3. Same as a normal backup really. The .env and the db backup arer critical

yeah the prune didn't gave him much extra data

the main issue is we would have to take all the info straight from the vm as its prity much ''full'' then run all the commands on the dev machine to backup and migrate i suppose ?

Just to be clear you are not restoring to the same machine, but migrating to one with a bigger disk?

correct

Then yeah that is the best approach. It may be interesting trying to take the DB backup while the disk is full though

yeah thats the headscratcher

It is possible over ssh.. The other thing is just dont decomission the old server till the new one is up and running. But have the containers stopped except the db so that no jobs are processed

Also if you are curious on where the data is you can run the following command in the sql shell to see

sql
SELECT 
     table_schema as `Database`, 
     table_name AS `Table`, 
     round(((data_length + index_length) / 1024 / 1024), 2) `Size in MB` 
FROM information_schema.TABLES 
ORDER BY (data_length + index_length) DESC;

My guess is it will be either characterwalletjournals as the largest or character_notifications. Maybe corp wallet journals up there too

If you want I can prototype a remote backup command for you

would be nice

i'm just trying to help him out but if there are just that many transaction data and its legit that big he might have to bite the bullet and just upgrade the storage

Plenty of cpu on the old server?

it had 6 cores and was at 2%

so don't think that was the issue

it was 6cores 6gb ram 60gb hdd

I mean the other thing is if thats where it is and he doesnt care about old data then we can manually remove old transactions easily

think they wanna preserve some old data as some tokens are now invalid due to inactivity but they ''might'' come back

Rog

More thinking about for backup trading CPU for network speed

where its going is fine

idea is to spin up a dev on that one migrate it to v5 and then return it back to the vm (if its still 55gb well then he has to decide to grow the vm at his vps or not)

It wont magically shrink....

hmhm

I do think the maintenaince job isn't running since it doesn't have enough time to run

Not sure what you mean by doesnt have enough time?

It can be manually run docker-compose exec seat-worker php artisan seat:admin:maintenance

the docker just stops it feels like

oh

ill get him to run that

As for the remote backup command, you can do something like this

bash
ssh username@seat-server 'docker compose -f /opt/seat-docker/docker-compose.yml exec mariadb mysqldump "seat" -u"DB_USER" -p"DB_PASS"' | gzip > seatv4_backup.sql.gz

Its not tested but should be a good starting point. An improvement would be running gzip on the server side but I coulnt test how the sytax would look

Also in general a helpful cli tool I have foound for tracing what is using all the disk space is ncdu

gotcha i think he's at work (euro tz) but ill start fiddling around with him when he gets home

Enjoy 🙂 I am off to bed so wont be able to help real time

all good think we can get somewhere with this and i suppose there are more people awake at that time goodnight!

You might be able to just run mysqldump as a remote operator not local to the host.

/dev/vda1 79G 46G 29G 62% / This is 4000+ tokens over a fair amount of time - with some prolific traders 🙂

we just ended streaming and need to get some groceries he ran the maintaince command but sees we need to do some digging more

hmm

Thought... Does he have the append only option set on the redis container?

If so I works disable that then restart redis

Did they pivot away from the redis container configs at all?

not sure astral have to check with him, had some other things yesterday so no time for it yet will dm him to see if/when he can do things

well he's on a road trip till sunday so guess we wil have to wait till then,,,

I like how jita 4-4 is shown as "unknown" .. hm

Seized by Drifter activity ⚔️

I have at least a few Citadels that have been renamed and still have their original name in SeAT (yes, plenty of tokens have access to the structure so it’s not an issue there). Wonder if it could be related to why stations remain unknown as I’ve seen similar behavior with citadels showing unknown where at least one user token can definitely query for the proper info. Maybe it’s based on the token batch that attempts to pull the citadel info, like if a character without access has an asset there or the API call fails.

more than likely

Citadel names are only fetched once, so it makes sense that it doesn't pick up the new name.

regarding the structure that doesn't load: Roughly, what is the ratio of characters that can access the structure vs characters that can't?

That would definitely explain it. The corp structures tab does show it, but I haven't looked at how those are fetched or where they're stored.

You can delete it from the universe_structures table and it should reload it.

I'd say it's likely to be 60:40-ish.

60 being the tokens that could access it.

I'd have to do some SQL fun to understand which of those characters have an asset in the structure(s) that might not have access to it, as I also assume there's a component there where it's not just trying random tokens.

hmm. I just noticed that per structure, it only allows one character to make a try at resolving names per hour. If it always picks a character that cannot access it, it will never load. But with 60/40, it should pick a character that can access the structure at some point.

I'd like to modify it to be one try per character, but I'm a bit afraid of the impact on the error limit if every character tries to fetch every unknown structure with an empty access cache

@user_131158012831203328 Are your citadel problems when starting the stack still this bad? I've been thinking about having an option to store the citadel access cache in the database. Performance-wise, it should mostly read to more reads

I'll have to see if I can find a good example of it, it is certainly possible it has self-resolved since that time, it has been a little while at least.

I have folk that drop an item in every citadel they come across in space

I have a lot of those errors

It looks like a majority of them have self-resolved over time as you expected, I admit I hadn't looked for a while at it. I'll have to dig into the few floating around - but those may be just due to legitimate access issues. I'd potentially assume it could be compounded by a large number of tokens that don't have access to the citadel but have assets in it? e.g. if you have 50 character tokens that don't have access to a structure but have at least one asset in it, and you have 5 tokens that do have access to the structure with an asset. It may also partially be due to the rate limiting of citadel jobs as they all stack up shortly after startup and the laravel jobs time out waiting in the queue when there are a large number of citadels to query - but that behavior is not something I'd bet on knowing all the nuances to 🙂

starting the stack?

my seat server has been running for 26 days, it's sitting at 413k errors past 7 days

mines at 1800 errors , but its SLOW

not sure why its so slow

Out of memory/disk?

Hello . need help . I installed everything according to the instructions, but when I go in I get . 502 Bad Gateway nginx . We need someone who understands this. Ubuntu 20.04 + Plesk

raw on the server?

not docker?

not docker . Страница недоступна Сайт seat.devit.lv HTTP ERROR 500 2025-02-01 21:42:12 Error 404 HEAD /auth/eve/callback HTTP/1.0 curl/7.68.0 677 Доступ к Apache SSL/TLS

Did you do this?

https://eveseat.github.io/docs/configuration/esi_configuration/

everything is already installed. Now the question is why this pops up?

That error looks like it's from CCP.. I would double check the call back URL is correct.

@user_528579767562403851 i think that is your issue ^^^

return link? or "seat" or port? 🙂 the port is correct because I have nginx and Plesk Obsidian

is the redirect URL in your .env exactly the same as on the eve developer portal?

.docs.sso

@user_614098468218339348, https://eveseat.github.io/docs/configuration/esi_configuration/

yes / callback link one and the same

I think redirect URL for SeAT must be: your.domain/auth/eve/callback

any reason you are proxy_passing the redirect url in your conf?

for reference, there is the recommended nginx configuration for seat: https://eveseat.github.io/docs/installation/manual_installation/#nginx-configuration

with this config

is there a person who can set it up completely? for PLEX or real money.

so if I understand it right, with the first config, the seat start page that redirects you to the CCP login page works, but when CCP redirects you back to seat, you get an error?

Hello i am getting serveral failed jobs for industry and im not sure why. I do not have any industry plugins installed.

yes / callback not work

try removing the location /auth/eve/callback part. As far as I understand, it just redirects to the same URL as originally

According to the document that is also incorrect..?? the docuementation shows it as

root /var/www/seat/public;

just double checking if his path is correct

I mean if the seat files are there, I guess it is right

you don't just accidentally type that

correct

So there are no people willing to earn some extra money? 🙂

@user_528579767562403851 there are many parts to what you are asking. Are you going to buy the domain name, and rent the servers? Will you then just give me your login (not recommended). Are you going to buy the cheapest server out there and complain when it doesn't work. And it isn't a simple "set this up for me" are you expecting a service contract or just ping and pay every other day when stuff updates. If it breaks in 2 months what will you do? I don't want to come behind someone else.

heh

define "some"

(not that I am offering)

😛

we have been continuing to help, even after offering payment.

Right now, it's your turn though. What happens if you remove the redirect loop?

I have a hosting company. My servers are personal. There is someone to service them. I also have a domain name. On Ubuntu 20.04 there is Plesk Obsidian. This is the difficulty, to combine the site with the hosting system. I found a way out and now I am raising KVM